Professional Documents
Culture Documents
RC wall
RC floor slab
Starter-bars
fixed to
Lap leng
wall rebar th
Wall rebar
36
Gerber JD, Van Zijl GPAG. Alternative wall-to-slab connection systems in reinforced concrete structures.
J. S. Afr. Inst. Civ. Eng. 2017:59(3), Art. #1633, 12 pages. http://dx.doi.org/10.17159/2309-8775/2017/v59n3a5
are designed to increase the construction
speed and efficiency, while minimising the
cost of labour and material (Rupasinghe &
Nolan 2007). The sliding nature of these
90°
systems requires that no elements may 90°
be protruding from the wall while under
construction. This requirement prohibits the 90°
use of the conventional system where starter-
bars are cast directly into the wall and fixed
90°
through the formwork, as seen in Figure 1.
The result of this limitation is that alterna-
tive methods are used to ensure rebar con-
tinuity between the floor and wall, without Figure 2 B
end-out bar shape: site-installed (reworked from Ancon Building Products 2011)
compromising the operation of the sliding
formwork. Alternative systems are also used
in cantilever balconies or walkways, or any Shape code 99 starter-bar
structure where the wall is cast ahead of the
connecting floor.
Currently, the most commonly used Dimpled steel casing
method in South Africa to ensure this con-
tinuity is a rebar bend-out system. Rebar is
pre-bent and fixed to the wall reinforcement,
and only straightened after the formwork is
removed. The shape of the pre-bent bars is
shown in Figure 2. This system is relatively
simple to install and can be assembled on site
using regular rebar ordered from a bending
yard. Although this system does not require Steel lid
highly skilled workers, the installation process
can still be tedious and time-consuming. Each Straightened rebar. Only done after
box is cast inside the wall
starter-bar has to be securely placed at the
correct position in the wall in order to line up
with the connecting floor, while individual
holes need to be drilled in the formwork at the Figure 3 B
end-out starter-bar unit (reworked from Ancon Building Products 2011)
position of each starter-bar leg.
In order to make the installation process
of the bent-out bars easier and faster, a Steel anchor with thread
modified approach is adopted in which the
pre-bent rebar is placed inside a steel casing,
as illustrated in Figure 3. This modification
improves the ease and speed of installation,
but in turn makes the system more expen-
sive. Another drawback is that only a few
companies locally manufacture this system,
and it is therefore not readily available Temporary
throughout South Africa. screw to keep Timber carrier
In the case of the bend-out systems, plas- anchor in
place
tic deformation of the steel is unavoidable
during the cold-bending and straightening
of the rebar. Furthermore, the mechani-
cal properties of rebar are more prone to Figure 4 C
ast-in anchors, pre-assembled in timber carriers (reworked from Ancon Building
changes during cold-bending. This change Products 2014)
in mechanical properties has been observed
to cause a reduction in the tensile yield the steel casing has on the bond strength of anchor is one example of such a mechanical
strength, modulus of elasticity and even the concrete connection, as it is permanently splice, as shown in Figure 4.
the ultimate strength of the steel (Chun & cast inside the structure. The failure mode for direct tensile
Ha 2014). Currently there is not enough The only reference in South African codes anchors is based on a model with a breakout
information available for engineers to to bend-out bars is found in SABS 10144, prism with an angle of approximately 35°
accurately take this effect into account when where it is recommended that the size of as shown in Figure 5. This can roughly
designing a connection. the bend-out bars is limited to Y10 or R16, be translated to a cone shape with a base
Considering the modified bend-out and that mechanical splices should be used equal to three times the effective anchor
system with the rebar encased in a steel box, if greater strength is required (SANS 2012). embedment depth, i.e. 3 × heff (Cairns 2010).
further uncertainty arises as to what effect Rebar couplers connected to a cast-in steel This pull-out model is referred to as a cone
Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017 37
RC wall
RC wall
heff
Tensile
force ‘Modified’ cone shape
heff RC floor
Cast-in anchor
Tensile force
1.5heff
Cast-in
z Coupled moment
anchor d
Tensile
force ‘Full’ cone boundary Compressive force
1.5heff 1.5heff
Approximately
35°
Figure 5 C
one failure for cast-in anchor in
direct tension (reworked from Ancon
Building Products 2014) Figure 6 M
odified cone failure for cast-in anchor in moment connections (reworked from Ancon
Building Products 2014)
failure mechanism. The design procedures
for anchors in direct tension are well esta to the relatively high price of this system, its The investigation includes a study of each
blished and documented in several design use is usually limited to projects where struc- individual system in order to gain an under-
codes, such as the New Zealand design tural alterations or extensions necessitate the standing of the critical aspects and para
codes (NZS 2006) and American Concrete anchoring of rebar in existing concrete. meters that will affect their structural per-
Institute codes (ACI 2008). In the case of epoxy-based adhesive formance. Preliminary tests are conducted to
These existing procedures, however, do dowels, the performance and installation quantify these aspects, followed by full-scale
not cover anchors within moment-resisting procedures are well researched and docu- testing and finite element modelling (FEM)
connections, such as wall-to-slab applica- mented in the reference material provided by to confirm any local phenomena observed
tions. A recent research investigation at the manufacturers. The structural engineer with the preliminary tests.
Heriot Watt University tested moment con- has enough information to properly design A comparison is drawn between the prac-
nections using cast-in anchors to determine a connection and prescribe the necessary tical aspects of the systems, including ease
the degree of enhancement in concrete installation procedures to ultimately have and time of installation, cost, quality control
pull-out capacity, and to establish a design enough confidence that the connection and availability of products.
method based on the results (Cairns 2010). will perform adequately. With most of the The main objective of this study is to pro-
In moment connections, such as wall- other advanced methods this information is vide the engineer with recommendations on
to-slab joints, the bottom portion of the not available. how to choose the most appropriate system
slab will create a compression block in From the above it is evident that there for a wall-to-slab moment connection and
close proximity to the bearing surface of are a number of different systems available conduct the design to be as cost-effective as
the anchor. The transmission of the force in reinforced concrete wall-to-slab connec- possible, without compromising the margin
between the bearing face and the compres- tions. The challenge for the design engineer of safety as defined in typical design codes.
sion force, forming part of the moment remains to choose the most suitable system Secondly, construction guidelines are
couple in the slab, will tend to be taken by a for the specific application at hand. In to be provided for safe application of these
direct compression strut rather than by shear order to make an informed decision, more systems on site, in order to comply with the
or tension. This will decrease the break-out information is needed to understand how designed connection.
plane resisted by tension and will therefore these systems perform structurally and
create a modified cone shape, as illustrated what the aspects to consider are when
in Figure 6. As concrete is stronger in com- designing them. EXPERIMENTAL DESIGN
pression than tension, the cone pull-out The focus of this research is to investigate
resistance will be increased. The tests also how different types of connection systems Full-scale tests
indicated that the enhancement is greatly compare to the conventional system. The In order to compare the different connection
impacted by the ratio of the depth of anchor research investigation includes the com- systems within the time constraint of this
head embedment (heff ) to the effective depth parison between four different wall-to-slab research, a best practice approach was adopt-
of the anchor in the slab (d), the specific connection systems, namely: ed in the physical experimental phase. Four
ratio being heff /d (Cairns 2010). ■■ Continuous starter-bar (Model A) full-scale models of the different connec-
Epoxy-based adhesive dowels are also ■■ Site-installed bend-out bar (Model B) tion systems were built in the Stellenbosch
widely used in South Africa to anchor rebar in ■■ Pre-assembled bend-out bar (Model C) University Structural Laboratory. The
already cast concrete elements. However, due ■■ Cast-in anchor (Model D) models represent a wall-to-slab moment
38 Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017
than the nominal bar diameter to ensure that
300
the effective area is not compromised. The
anchor length was 115 mm, with a 40 mm
diameter head and an embedment depth of
250
142 mm.
Fixed In order to simulate practice as closely as
possible and to ensure continuity, a typical
reinforcement configuration for all the wall
sections was used. The vertical reinforce-
ment was Y12 bars spaced at 150 mm (c/c)
Applied load on both sides of the wall. Similarly, the
horizontal reinforcement was Y12 bars
1 700
200
spaced at 200 mm c/c. This configuration
complies with the detailing requirements of
250
Specimen
combination of 19 mm and 13 mm Hornfells
(1 000 wide)
stone was used as aggregate, and the sand
Fixed component consisted of both dune sand and
crusher dust. The concrete was delivered to
the laboratory in two separate batches – the
All dimensions in millimetres Fixed first batch was used to cast the walls for all
four models, and seven days later the second
Figure 7 F ull-scale specimen with boundary conditions and applied load batch was used for the cantilever floors.
The concrete compressive strength was
connection found in a typical high rise for the variable actions. The procedure can determined by casting and curing 100 mm
office block. be summarised through the following steps: cubes from the concrete batches for the
The models were all identical, with a ■■ Step 1: Determine forces acting on the walls and the slabs. The cubes were stripped
250 mm thick cantilever floor connected to structure. after one day and cured under the same
a 300 mm thick wall. The systems were all ■■ Step 2: Calculate possible load conditions in the laboratory. On the day of
designed to have the same moment and shear combinations. the wall-to-slab connection tests, the cubes
capacity, based on the actions associated ■■ Step 3: Analyse structure to determine were tested in a Contest 1 MN materials
with a typical office environment. design forces and moments at ULS. testing machine. Extra cubes were prepared,
When designing RC structures, the ■■ Step 4: Design section for flexure. and tested at the age of seven days to moni-
ultimate limit state (ULS) design usu- ■■ Step 5: Design section for shear. tor the strength development, to ensure a
ally governs (Robberts & Marshall 2010). ■■ Step 6: Choose rebar configuration (size high likelihood that the required 40 MPa
Therefore the design procedure for the and spacing). strength would be attained at the time of the
conditions at ULS, and the principles as set ■■ Step 7: Check anchorage and lap lengths. actual wall-to-slab tests. A set of three cubes
out in SANS 10100-1:2000 (SANS 2000), ■■ Step 8: Detail connection reinforcement. representing the walls and the slabs respec-
were followed to design the connection All the calculations for the models that tively were tested on the day of the actual
between the wall and the slab in the full- contained the bend-out systems were based wall-to-slab tests, with an average compres-
scale models. Detailed design calculations on the assumption that the rebar had already sive strength of 40.5 MPa and coefficient of
were performed for a typical office building been straightened, and therefore reflected variation of 0.011 for the walls, and 40.8 MPa
to ensure relevance of the chosen specimen the conventional system. The only differ- (CoV 0.009) for the slabs.
geometry and loading position. The loading ence was in the detailing of the starter-bars. After 28 days the models were individu-
position caused a realistic combination of The design procedure for the cast-in anchor ally tested in a special steel frame designed
internal shear and bending moment at the system followed the approach as set out in to ensure that the desired boundary condi-
wall-to-slab connection. For the ULS the the CEB Design Guide and making use of the tions are met. An illustration of the applied
load was 49.5 kN and for the SLS service- modified cone behaviour (CEB 1997). loads, boundary conditions and a full-scale
ability limit state (SLS) it was 40.5 kN. The reinforcement configuration that specimen, is presented in Figure 7, with the
These values are based on a variable action satisfied the connection design for Models test configuration in Figure 8.
of 3 kPa and additional permanent actions A, B and C comprised six Y12 starter-bars During the experimental testing a load
equal to 3.5 kPa to account for the weight spaced at 150 mm centre-to-centre (c/c), cell was used to capture the applied load,
of screed, ceiling, lighting and air-condi- with 25 mm cover. The cast-in anchor while seven HBM linear variable differential
tioning equipment. For the ULS, the partial system installed in Model D had a similar transformers (LVDTs) were used to measure
factor for the variable actions was taken as arrangement, with six anchors also spaced at the structural response of the specimens.
1.6, while 1.2 was used for the permanent 150 c/c. The anchors were supplied by Ancon The LVDTs were located at specific points
actions, including the self-weight of the Building Products. The continuation bars along the specimens in order to calculate
slab. The partial factors used for the SLS, were also 12 mm in diameter, each with a the relative wall-to-slab displacement. An
were 1.1 for the permanent actions and 1.0 threaded end, the thread being a size larger additional LVDT was located at the back
Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017 39
of specimens. Significant care was taken
to avoid geometrical imperfection in the
specimen dimensions and in the positioning
A-frame
Instron of the steel reinforcing bars and the con-
nection systems. The fresh concrete showed
Specimen tied to A-frame no segregation, nor leakage from the form-
work. The low variability in concrete cube
A-frame fixed
strengths reported above indicated good
to laboratory control. The actual stress-strain behaviour
strong floor of the steel bars was carefully measured, as
reported in subsequent sections of this paper.
The authors postulate that, given the care
Specimen in controlling the material and geometry of
the specimens, a reasonable reflection of the
various connection systems is given by the
results of the tests on a single specimen of
each type. Variability in the supplied connec-
Instron head
tion systems is not reflected in the results,
and will be tested in a next phase.
Ball-bearing
Steel bars tensile tests
154 × 154 × 35 In order to understand the effect of cold-
H-section spreader beam
bending and straightening on the material
properties of reinforcement, a series of
Specimen tensile tests were also conducted. The
specimens included a set of Y10 and Y12
Figure 8: Full-scale test configuration deformed bars. These are typical diameters
used in wall-to-slab connections, as larger
diameters are not suitable for cold-bending
on site. Half of each set was tested as straight
bars, while the remaining bars were cold-
bent and straightened after two weeks before
being tested.
Specimen All the tensile tests were performed
with a Zwick Universal Testing Machine
and complied with the specifications of
SANS 6892-1:2010 (SANS 2010).
LVDT clamp
Both sets of Y12 and Y10 rebar were
ordered from a local bending yard. The
reinforcement was locally manufactured
at Arcelor Mittal, under the registered
Top of bracket Machined portion trade name NOSTRAR, to comply with
of specimen the specifications of SANS 920:2011
(SANS 2011) for 450 MPa deformed
reinforcement bars.
3x LVDTs All the bars in a set were cut from the
same batch of steel. Half of the bars in each
set were bent to a 90° angle. The bars were
3x Bracket spacers Bottom of bracket bent with a standard bending table, and the
bending radius complied with the values as
prescribed in SANS 920:2011 (SANS 2011).
The bent bars were all straightened after
Figure 9: Tensile testing LVDT set-up 14 days, simulating the usual time lapse
between the bending of the rebar and the
of the specimens to record the horizontal The results were presented through load straightening on site. The bars were placed
movement of the A-frame at a height of displacement curves and crack patterns in a table vice clamp and straightened using
525 mm. observed at applied loads associated with the a pipe as lever arm. A constant force was
A standard concrete crack width ruler was SLS, ULS and the load at failure. applied in a single motion until the bar was
used to measure the width of all visible cracks Due to the size and complexity, one reasonably straight. In some cases the bars
at three different stages during the tests. An specimen of each wall slab connection type were further clamped and straightened,
Aramis camera was also installed to con- was tested. To address the inherent vari- within a tolerance of 3°, to ensure they are
tinuously track the crack development in the ability in reinforced concrete behaviour, best straight enough to fit between the cross
vicinity of the wall-to-slab joint (GOM 2003). practice was followed in the manufacturing heads of the testing machine.
40 Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017
Stress-strain curve
Wall Wall
Stress
fy
E Floor Floor
1
Strain
Figure 10: D
etermination of yield stress – Model A Model B
Continuous starter-bar system Site-installed bend-out bar system
direct method
Stress-strain curve
0.2% offset line Wall Wall
Stress
fy
Floor Floor
E
1
0.002
Strain
Model C Model D
Pre-assembled bend-out bar system Cast-in anchor system
Figure 11: D
etermination of yield stress –
offset method Figure 12: Cracks at Stage 1 – wall joint
It should be noted that, in order to During the tensile testing the applied tensile EXPERIMENTAL RESULTS
apply the principle of best practice, the force was measured with the load cell of the
bars were not hammered or tampered testing machine, while the extension over the Full-scale test results
with during the straightening process. It is gauge length of the specimens was recorded Figure 12 presents the crack patterns at
acknowledged that, in practice, some ham- by three LVDTs, as shown in Figure 9. By Stage 1 of the tests (SLS). Photos captured
mering may occur on site, but this was not taking the average reading between the three by the Aramis camera are superimposed
simulated in this experimental investigation. LVDTs any eccentricity in the alignment of with rendered images indicating the strains
The results are, therefore, a comparison the specimen was taken care of. The record- measured on the specimens. From these ren-
between the material properties under best ing frequency was 10 Hz. dered images the crack patterns are clearly
practice conditions. The methods as set out in SANS 68921:2010 visible. The location of the major cracks in
A summary of the specimen sets are pre- (SANS 2010) were used to determine the all four systems corresponds to the position
sented in Table 1, with the abbreviations that material properties of each specimen. The and shape of the construction joint of the
will be used hereafter. E-modulus (E) was determined by first plotting particular system used. All the specimens
a trend line for the data in the linear portion of recorded one major crack of between 0.1 mm
Table 1 A
bbreviations for tensile testing the stress-strain graph, and then using the gra- and 0.4 mm, with Model A recording the
specimen sets dient of that line as the value for E (SANS 2010). widest crack of 0.4 mm. This crack was
The yield stress (fy) was determined by one of observed on the rear side of the wall and is
Preliminary tensile testing
two possible methods, based on the shape of therefore not visible in Figure 12. Apart from
Y10 straight bars (machined) Y10-S the stress-strain curve. When the curve had this crack, all the other cracks fell within
Y10 bent and straightened bars Y10-B a definite linear portion up to a certain point the general limit of 0.3 mm for structures
and then flattened, as seen in Figure 10, the exposed to a serviceability load (SANS 2000).
Y12 straight bars (machined) Y12-S
yield stress (fy) was taken as the stress value Once the tests went past Stage 2 (ultimate
Y12 bent bars Y12-B corresponding to that point on the graph. If, limit state) the crack development in the
however, the stress-strain curve had no definite first three systems was similar, both in size
Full-scale model rebars
point at which it flattened and the gradient for and pattern, and was clearly indicative of a
Wall bars Y12-WR the linear portion just gradually reduced to cre- connection where the rebar was starting to
Model A starter-bars (machined) Y12-FRA ate a curved graph, then the offset method was yield. In contrast to this, the pattern that
used as seen in Figure 11. was observed in Model D strongly sug-
Model B starter-bars Y12-FRB
The ultimate stress ( fu) was recorded gested a cone pull-out failure. Looking at the
Model C starter-bars Y12-FRC as the highest applied load during the test photos in Figure 13, taken at the final stage
divided by the original cross-sectional area of the tests, the cone-shaped crack pattern
Model D starter-bars (machined) Y12-FRD
of the rebar. in Model D is clearly visible. Although the
Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017 41
after which there is a sudden drop in the
load. This drop in resistance is followed by a
Wall Wall gradual softening response. Consequently it
has a significantly lower ultimate resistance
and the earliest onset of deflection softening
of the four models. This behaviour is associ-
Floor Floor ated with a pull-out cone failure mechanism.
Table 2 presents a summary of the
displacements and rotations recorded at the
first two stages of the tests.
B 3.2 1.6
C 4.6 2.4
D 3.6 2.3
Floor Floor
Displacement at ULS (Stage 2)
A 3.8 2.2
B 4.8 2.4
40
Table 3 Ultimate capacity summary
Specimen A
Specimen B Specimen number Applied load (kN)
20
Specimen C
A 98.12
Specimen D
0
B 93.56
0 20 40 60 80 100
Displacement (mm) C 87.86
Specimen A Specimen B Specimen C Specimen D
D 75.86
42 Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017
The average yield strength ( fy) for the
Stress-strain curve
Y10-S bars was 550 MPa, which is well above
700
the characteristic value of 450 MPa used in
600
design codes. However, according to the Joint
Committee on Structural Safety (JCSS 2001),
500 the average strength of rebar is expected to
be around 2 standard deviations above the
Stress (MPa)
Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017 43
Table 6 Finite element model summary
Element types
Mesh configuration
Concrete 50 × 50 × 50 mm elements
Reinforcement Automated
Material models
Compressive behaviour –
Thorenfeldt (a) (b)
translation
Load conditions
Convergence Figure 17: Comparison of experimental data with FEA data of Model A
Energy – 0.001 tolerance
criteria
44 Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017
Crack width (m)
+4.07928e-004
0.0%
+3.82432e-004
Wall Wall 0.1%
+3.56937e-004
0.1%
+3.31441e-004
0.1%
+3.05946e-004
0.1%
+2.80450e-004
0.4
0.1%
+2.54955e-004
0.6%
Floor +2.29459e-004
Floor 0.6%
+2.03964e-004
1.1%
+1.78468e-004
1.3%
+1.52973e-004
2.0%
+1.27477e-004
2.2%
+1.01982e-004
2.3%
+7.64864e-005
3.9%
+5.09910e-005
Physical experiment Numerical analysis 5.3%
+2.54955e-005
Crack pattern and width at SLS Crack pattern and width at SLS 80.0%
+0.00000e+000
0.3%
+2.91425e-004
Floor Floor 0.6%
+2.59044e-004
0.9%
+2.26664e-004
1.2%
+1.94283e-004
0. 3
2.0%
+1.61903e-004
2.1%
+1.29522e-004
2.3%
+9.71417e-005
3.9%
+6.47611e-005
5.5%
Physical experiment Numerical analysis +3.23806e-005
Crack pattern and width at ULS Crack pattern and width at ULS 80.5%
+0.00000e+000
Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017 45
for this system is three times that of the pre-
Cantilever end displacement –Y12-S rebar
assembled bend-out system, as presented in
100
Table 7.
90 Table 7 presents a summary of the
80
material costs for the respective systems,
determined in August 2016. Both the local
70 rates and the rates in the United Kingdom
Applied load (kN)
30
Table 7 Combined material cost comparison
20 Rate Rate
System (UK) (RSA)
10 (£/m) (R/m)
Figure 20: Results for finite element analyses – S-series models Cast-in anchor 83 *
46 Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017
performance of the site-installed system that they have not been damaged in any Available at: http://www.ancon.co.uk (accessed in
is slightly superior and the material costs way during the straightening process. June 2016).
47% less, practical considerations and ■■ Any reinforcement that has been Cairns, J 2010. Design analysis for moment connections
project-specific aspects might be of greater cold-bent past 45° should be consid- with KS Threaded Anchors. Technical Report
influence when choosing the most desir- ered to have a reduced capacity and (unpublished). Edinburgh: Heriot Watt University.
able system. It is therefore recommended ductility, and the design of the specific CEB 1997. Design of Fastenings in Concrete. London:
that the suitability of a specific alternative element should be checked, using the Thomas Telford.
system should be evaluated against all the modified yield stress and E-modulus as Chun, S & Ha, T 2014. Cyclic behavior of wall-slab
aspects for any given project. recommended above. joints with lap splices of cold- straightened rebars
■■ If one of the bend-out systems is chosen, ■■ The use of cast-in anchors could be a and mechanical splices. Journal of Structural
the design should be conducted with a feasible option in highly congested walls Engineering [Online], 141(2).
modified yield stress and E-modulus for or when larger moment capacities are Deaton, J B 2013. Nonlinear finite element analysis of
the flexural reinforcement in the joint. The required. The enhancing effect of the reinforced concrete exterior beam-column joints
recommended value for the yield stress modified cone behaviour can be adopted with nonseismic detailing. PhD thesis. Atlanta, GA:
is 0.8* f y and 150 GPa for the E-modulus. to produce more economical designs, but Georgia Institute of Technology.
The design engineer should also strive to once again larger deflections and cracks Diana 2014. Diana User’s Manual. Delft: TNO Diana.
keep the elongation of the reinforcement should be expected and accounted for. GOM 2003. GOM Aramis User Manual. 3.11 Software,
in the connection to a minimum. ■■ The manufacturer’s specifications and Vol. 49. Braunschweig, Germany: GOM.
■■ The size of the reinforcement to be used installation procedures should also be JCSS (Joint Committee on Structural Safety) 2001.
in any bend-out system should be limited strictly followed when using the cast-in Probabilistic Model Code for New and Existing
to a diameter of 12 mm. anchor system. A proper tightness check Structures, Part 3.2 Limit State Function. Available
■■ When using site-installed bend-out of all the continuation rebar should at: http://www.jcss.byg.dtu.dk/-/media/Subsites/.../
systems, good site supervision and quality be conducted, before the fixing of the probabilistic_model_code (accessed in June 2016).
control are necessary to ensure that the reinforcement for the floor commences. NZS (New Zealand Design Codes) 2006. NZS 3101.
starter-bars are fixed securely and at the The tightness can be checked by using a Concrete Structures Standard. Part 1: 2006. The
correct level. It is recommended that the standard hand wrench. Design of Concrete Structures. Wellington, NZ:
future floor level is clearly indicated on the ■■ The thread at the end of the continuation Standards New Zealand.
horizontal wall reinforcement to help iden- reinforcement, supplied by the manu- Robberts, J M & Marshall, V 2010. Analysis and design
tify starter-bars that are fixed out of place. facturer, should be a size larger than the of concrete structures. Johannesburg: Nuclear
■■ Care should be taken when chipping nominal diameter of the bar. Couplers Structural Engineering.
away the concrete surrounding the pre- with tapered threads should be avoided Rupasinghe, R & Nolan, É 2007. Formwork for Modern,
bent rebar in the site-installed bend-out as they only obtain their tensile capacity Efficient Concrete Construction. Garston-Watford,
system. The pre-bent rebar should first when they are fully screwed in. The con- UK: IHS BRE Press.
be fully exposed, with at least 30 mm stant diameter couplers, as used in this SANS (South African National Standard) 2000.
clear spacing behind the bend, before any study, gradually build up their capacity SANS 10100–1:2000. The Structural Use of Concrete,
form of straightening is considered. A with each turn of the continuation bar. Part 1: Design. Pretoria: SABS Standards Division.
steel pipe should be used to straighten the SANS (South African National Standard) 2010. SANS
bars. The pipe should be placed over the 68921–1:2010. Metallic materials – Tensile Testing.
reinforcement and continuously pushed REFERENCES Part 1: Method of Test at Room Temperature.
inwards as far as possible, while straight- ACI (American Concrete Institute) 2008. ACI 318-08: Pretoria: SABS Standards Division
ening the rebar in one smooth motion. If Building Code Requirements for Structural Concrete SANS (South African National Standard) 2011. SANS
the bar is not entirely straight after the and Commentary, Appendix D. Michigan, MI: ACI. 920:2011. Steel Bars for Concrete Reinforcement.
first attempt, not more than one further Ancon Building Products 2011. Reinforcement Pretoria: SABS Standards Division
adjustment should be considered. continuity systems for the construction industry – SANS (South African National Standard) 2012. SABS
■■ Once the starter-bars are straightened, the Eazistrip. Available at: http://www.ancon.co.uk 0144:1995. Detailing of Steel Reinforcement for
concrete at the joint should be properly (accessed in June 2016). Concrete, 2nd ed. Pretoria: SABS Standards Division.
scabbled to ensure aggregate interlock. All Ancon Building Products 2014. Reinforcement continuity Schreppers, G & Frissen, C & Kang, H 2011. Prediction of
starter-bars should be checked to verify systems for the construction industry – KSN Anchors. crack-width and crack-pattern. Delft: TNO Diana BV.
Journal of the South African Institution of Civil Engineering • Volume 59 Number 3 September 2017 47